Digital Hemispherical Photography based LAI estimate detects higher underestimation of Radiative Transfer Model extracted LAI in a mangrove forest site, India

A hybrid leaf area index (LAI) retrieval algorithm is embedded with Sentinel Application Platform (SNAP). Here Artificial Neural Network (ANN) works along with a one-dimensional Radiative Transfer Model (RTM), PROSAIL (PROSPECT+SAIL), where PROSPECT simulates the reflectance and transmittance as a function of leaf biochemistry and anatomical structure and SAIL simulates top of canopy reflectance. This retrieval system is called as SNAP biophysical processor or SL2P. Because of its ease of use and integration within the image processing software, using this processor, LAI was estimated in Bhitarkanika Wildlife Sanctuary (BWLS), the 2nd largest mangrove ecosystem of India and known as a Ramsar site. Field campaigns were conducted in March from 2019 to 2021, and 174 locations were covered in an Elementary Sampling Unit (ESU) scale (20m x 20m). From each ESUs, 13 Digital Hemispherical Photographs (DHP) were collected and processed in CAN-EYE open-source software. SL2P processed LAI from both Sentinel-2 (S2), and Landsat-8 (L8) were compared with DHP-extracted LAI. These comparisons revealed much higher RMSE (>2) between SL2P and DHP extracted LAI. An ANN-based regression model was developed to reduce the error, where DHP-based LAI was considered as a function of surface reflectance bands of S2 and L8. Results show that L8 performs comparatively better than S2 with respect to model fitting accuracy (RMSEL8 = 0.54 RMSES2 = 0.61) as well as validation accuracy (RMSEL8 = 0.61 and RMSES2 = 0.81). Still, the histograms of the estimated LAI rasters showed that LAIL8 is suppressed in the higher range of LAI, but LAIS2 is well distributed. Mangrove has a higher potential to sequester atmospheric CO2, so the correct representation of LAI is important to assess correct primary productivity. Continuous filed observation using DHP considering homogeneous and heterogeneous field condistion will offer to validate the existing global LAI products over mangroves of Eastern India.